Picosecond Reverse Saturable Absorption in King's Complex [ (C,H

We have demonstrated reverse saturable absorption of picosecond pulses at 532 nm using the organometallic tetramer cyclopentadienyliron carbonyl ...
0 downloads 0 Views 676KB Size
J. Phys. Chem. 1992,96,6313-6317 sufficient data are available the correlation factor is significant. It is shown that Lennard-Jonesparameters are not g o d correlating factors; the implications of this fact on the type of interaction potential effective for energy transfer from highly excited molecules are discussed.

Acknowledgment. This work was supported by the United States-Israel Binational Science Foundation and by the German-Israeli Foundation for the Promotion of Research.

References and Notes I I ) Tardv. D. C.: Rabinovitch. B. S . Chem. Rev. 1977. 77. 369. (2j Quaik, M.; Troe, J. Gas Ki’netic and Energy Transier, The Chemical Society: London, 1977; Vol. 2. ( 3 ) Oref, I.; Tardy, D. C. Chem. Rev. 1990, 90, 1407. (4) Krajnovitch, D. J.; Parmenter, C. S.;Catlett, D. L. Chem. Rev. 1987, 87. 237. - , -( 5 ) Lambert, J. D. Vibrational and Rotational Relaxation in Gases; Clarendon Ress: Oxford, 1977. (6) Flynn, G . W. Acc. Chem. Res. 1981,1, 334. (7) Volbe. M.: Johnston. H. S . J . Am. Chem. Soc. 1956. 78. 3903. (8j Cha’n,’S. C.; RabinAtch, B. S.;Bryant, J. T.; Spicer, L. D.; Fujimoto, T.; Lin, Y. N.; Pavlou, S.P. J . Phys. Chem. 1970, 74, 3610. (9) Phillips. L. F. J . Phys. Chem. 1990, 94, 7482. (10) Philippoz, J. M.; Monat, R.; van den Bergh, H. J. Chem. Phys. 1990, 93, 8676.

6313

(1 1) Russel, K. E.; Simons, J. Proc. R. SOC.London A 1953, 21 7, 271. (12) Gilbert, R. G.; Smith, S . C. Theory of Unimolecular and Recombination Reactions; Blackwell Scientific Publications: Oxford, 1990. (13) Hippler, H.;Troe, J.; Wendelken, H. J. J. Chem. Phys. 1983, 78, 6709. (14) Shi, J.; Barker, J. R. J. Chem. Phys. 1988, 88, 6219. (15) Hippler, H.; Troe, J.; Wendelken, H. J. J. Chem. Phys. 1983, 78, 6718. (16) Hildebrand, J. H.; Scott, R. L. The solubility ofnonelectrolyres, 3rd ed.; Reinhold Publishing Co.: New York, 1950. (17) Fried,V.;Hameka, H. F.; Blukis, V . Physical Chemistry; Macmillan Publishing Co.: New York, 1977. (18) Hippler, H.; Otto, B.; Troe, J. Ber. Bunsen-Ges. Phys. Chem. 1989, 93, 428. (19) Yerram, M. L.; Brenner, J. D.; King, K. D.; Barker, J. R. J . Phys. Chem. 1990, 94,6341. (20) Marcoux, P. J.; Setser, D. W. J . Phys. Chem. 1978, 82, 97. (21) Fowler, R.; Guggenheim, E. A. Statistical Thermodynamics; The University Press: Cambridge, U.K., 1956. (22) Lin, Y. N.; Rabinovitch, B. S.J. Phys. Chem. 1970, 74, 3151. (23) Herscovitz, 0.; Tzidoni, E.; Oref, I. Chem. Phys. 1982, 71, 221. (24) (a) Nordholm, S.;Freasier, B. C.; Jolly, D. L. Chem. Phys. 1977,25, 433. (b) Freasier, B. C.; Jolly, D. L.; Nordholm, S.Chem. Phys. 1978, 32, 161. 169. ( 2 5 ) Sukowski, U.; Seilmeier, A.; Elsaesser, T.; Fischer, S.F. J . Chem. Phys. 1990, 93, 4094. (26) Lim, K. F.; Gilbert, R. G . J . Phys. Chem. 1990, 94,77. (27) MacKerell. A. D.: Karulus, M. J . Phvs. Chem. 1991. 95. 10559. (28) Hoover, W. A. Annu. Rev. Phys. Chem. 1983, 34, 103.

Picosecond Reverse Saturable Absorption in King’s Complex [(C,H,)Fe(CO)], G.R. Man,* D. R. Labergerie, S. J. Rychnovsky, Thomas F. Boggess, Arthur L. Smirl, Center for Laser Science and Engineering, 124 AMRF, Oakdale Campus, The University of Iowa, Iowa City, Iowa 52240

and Lee Tutt Hughes Research Laboratories, 3011 Malibu Canyon Road, Malibu, California 90265 (Received: February 10,1992)

We have demonstrated reverse saturable absorption of picosecond pulses at 532 nm using the organometallic tetramer cyclopentadienyliron carbonyl [(C5H5)Fe(CO)],dissolved in methylene chloride. Time-resolved pumpprobe measurements and numerical simulations based on a five-level model indicate that reverse saturable absorption of ultrafast optical pulses is due to excited singlet absorption in this tetramer. We have determined the lifetime of the first excited singlet state to be 120 ps, and have measured a singlet excited-state capture cross-section of 8 X lo-’*cm2,twice that of the ground-state cross-section. In addition, we have placed bounds on the lifetime of the second excited singlet state and the branching ratio between the singlet and triplet systems.

I. Introduction The phenomena of reverse saturable absorption (RSA), as reported in 1967 by Giuliano and H a , ’ was observed while dyes for saturable absorption were being investigated. Giuliano et al. reported that, in contrast to saturable absorption, some compounds, such as sudanschwartz-B and indanthrone, exhibit an absorption that increases with increasing intensity (RSA). Physically, the process can occur when the absorption cross-section of an excited molecular state exceeds that of the ground state. As the optical excitation rate increases, more molecules are promoted to the excited state, thus causing the absorption to increase with fluence. Because this process involves electronic transitions, materials that exhibit RSA generally have an extremely fast response, unlike the photochromic materials used in commercial sunglasses which depend on slow geometric rearrangements. Although RSA has not been extensively studied since the early work of Giuliano and H e s , recent interest in the potential application of RSA to optical lin~iting~-~ and mode locking of solid-state lasers6 has renewed study in such materials. In this paper, we examine the picosecond photodynamicsof the organometalliccompound cyclopentadienyliron carbonyl tetramer

[(CsH5)Fe(CO)]4. This compound was first synthesized by R. B. King’ in 1966 and commonly bears his name. The compound possesses high photostability, thermal (1 60 “C) stability, and chemical stability and has been previously shown to exhibit RSA on nanosecond time scales.* In this case the RSA was attributed to triplet-triplet excited-state absorption. As we report below, RSA also occurs in this molecule on picosecond time scales, indicating that singletsinglet transitions are also involved. In order to quantify the photodynamics that determine the RSA, we have performed time-resolved nonlinear spectroscopy on King’s complex in methylene chloride solution. Specifically, by comparing the results of pumpprobe transmission measurements performed with 532-nmpicosecond pulses to a five-level rateequation model, we extract values for the singlet lifetime and the excited singlet-state absorption cross-section, and we determine bounds on the lifetime of the second excited singlet state and the product of the singlet-triplet crossing rate with the triplet-triplet absorption cross-section. In the following section we describe the preparation of King’s complex and the experimental arrangements for the picosecond measurements. In section 111, we describe the results of single

0022-3654/92/2096-63 13$03.00/0 0 1992 American Chemical Society

6314 The Journal of Physical Chemistry, Vol. 96, No. 15, 1992

Allan et al. 3.0~10' h " I

E

2.5

"0 I

zj

v

2.0

+I

1.5

Figure 1. Schematic of the experimentalsetup used for the time-resolved pump-probe. and nonlinear transmission measurements.

0

2 1.0 32 0.5 3

pulse transmission (induced absorption) and time-resolved pumpprobe transmission measurements. We then discuss the possible optically-induced transitions in King's complex that give rise to the RSA and develop a five-level model used to analyze our results in section IV. Finally, our concluding remarks are presented in section V.

II. Experimental Section The King's complex used in these experiments was prepared from cyclopentadienyliron dicarbonyl dimer, which was purchased from Strem Chemicals and was used as received. The cyclopentadienyliron carbonyl tetramer, King's complex, was synthesized from the cyclopentadienylirondicarbonyl dimer by the method of White and Cunninghamgand recrystallized twice from methylene chloride/hexane. All measurementsreported here were performed on solutions of King's complex dissolved in methylene chloride. The methylene chloride was oxygen degassed with nitrogen by a vigorous bubbling purge for at least 2 h prior to use. This precaution avoids possible quenching of triplet-state population by oxygen contamination. All solutions were subsequently handled anaerobically. The nonlinear optical properties were measured using an excitation wavelength of 532 nm and a pulse duration of 25 ps (fwhm intensity) in two distinctly different types of experiments: single-pulse transmission and time-resolved pumpprobe measurements. The picosecond pulses were produced by a frequencydoubled, actively and passively mode-locked, Nd:YAG laser. The laser operates in the TEMWtransverse mode and produces temporally newGaussian pulses of nominally 30-ps duration (fwhm) at 1.064 pm. The actual pulse width was determined on a shot-by-shot basis using second harmonic techniques. The picosecond pump-probe and transmission experimental arrangement is shown schematically in Figure 1. A single pulse is split into two (an intense pump and a much weaker probe), a variable delay is introduced into the probe path, and the two pulses are recombined at the sample cell at a small angle (1.2'). The pump was collimated, and the probe was focused to measured Gaussian spots of radii 280 and 30 pm, respectively, at the sample. Considerable care was taken to ensure accurate overlap within the sample. The small angle between the beams and the fact that the probe spot size is considerably smaller than the pump ensure that the probe interrogates a radially uniform region of material in the 2-mm cell. The probe polarization, which was controlled by a half-wave plate and polarizer, was for most measurements selected to be perpendicular to the polarization of the pump. Absolute timing between the pump and probe (i.e., zero delay) was determined by translating a thin wafer of ZnSe, which is a twephoton absorber at 532 nm, into the beam interaction volume and measuring the probe transmission as a function of the relative delay between the pump and probe. The instantaneous twophoton-induced absorption of the probe is negligible except when the pump and probe are temporally and spatially coincident, and it is maximized at zero delay. Zero delay is, therefore, the delay at which the probe transmission is a minimum. We characterize the nonlinear absorption in King's complex by measuring the transmitted probe energy as a function of pump fluence and relative delay between the pump and probe. The single-pulse nonlinear transmission measurements were conducted by simply blocking single-pulse nonlinear transmission measurements were conduted by simply blocking the incident probe and measuring

wx

0.0

Wavelength (nm) Figure 2. Linear absorption spectrum of a 7 X M solution of King's complex in methylene chloride with solvent background subtracted. The extinction coefficient is 1035 cm-I M-l at 532 nm. 500

I

,

,

,

,

, ,

Input Fluence (mJ.cm-') Figure 3. Transmitted fluence as a function of input fluence in King's complex for picosecond pulses. The solid line represents linear transmission of -70%; the squares are the experimental points which clearly show the onset of reveme saturable absorption, Le., increasing absorption with fluence.

the pump transmission as a function.

III. Results and Discussion The ground-state absorption spectrum of a -7 X lo4 M methylene chloride solution of the cyclopentadienylironcarbonyl tetramer was measured in a 1-cm-optical-path quartz cell using a standard UV/visible spectrometer and is shown in Figure 2. This spectrum is background corrected, Le., the absorption due to the solvent alone has been removed, and it is consistent with thme reported in the literature.l0 For wavelengths between approximately 550 and lo00 nm, the spectrum consists of a broad, relatively weak absorption band. Upon close inspection this feature is Seen to be composed of two separate peaks, one centered near 650 nm and one centered near 775 nm. These peaks have been assigned to metal-metal core transitions.12 The 532-nm excitation used for the present experiments lies on the high energy side of these resonances and below the strong chargetransfer transitions that dominate the spectrum in the blue. The extinction coefficient at 532 nm is 1035 cm-l M-' companding to an optical absorption cm2. cross-section, uol,of 4.1 X The results of the single-pulse picosecond transmission measurement are shown in Figure 3, where we plot the transmitted fluence as a function of input fluence. For this measurement, the concentration of the solution was set such that the 2-mm cell had a small signal transmission of -70% at 532 nm, corresponding P M. The solid line represents the linear to a molarity of -9 X 1 response of the system. Comparison of the data with this line clearly shows the onset of RSA for input fluences in excess of 200

The Journal of Physical Chemistry, Vol. 96, No. IS, 1992 6315

Reverse Saturable Absorption in [(C5Hs)Fe(C0)l4

-8 a

2

0.00

-

-0.05

-

.r(

.il

-

-0.10

2

Pump Fluence = 1 1 mJ.cm-' 1

b

XPump Fluence = 28 mJ.cm-*:

'Ea -0.15 3

+Pump Fluence = 64 mJ.cm-'

B -0.20

-

5

e -0.25 ~

, , , , , , , , , , , 1 , , , 1 , , , 1 , , , ~

-200

0

200

400

600

000

1000

Probe Delay (ps) Figure 4. Timeresolved differential transmission measurementsof King's complex in methylene chloride using pulses of 532 nm with a fwhm of 25 ps. The linear transmission is -35%. A negative value for differential transmission indicates increased absorption. The solid lines are the numerical fits for a single set of parameters at the indicated fluence; see section IV and Table I.

mJ/cm2. Inspection of the transmitted beam profile showed that no significant self-lensing occurred as the beam passed through the cell, indicating that the system response was due to induced absorption. We emphasize that this optically-induced absorption occurs during the 25-ps optical pulse. Since we expect no significant intersystem crossing to occur on such time scales, we attribute the observed picosecond RSA to excited-state absorption within the singlet system. If the lifetime of the excited singlet state is on the order of or shorter than the optical pulse duration, this singlet excited-state absorption will depend on both the singlet lifetime, q0,and the absorption cross-section, u12. In such a case, we cannot isolate and separately measure these two quantities from the single-pulse transmission results. This problem can be overcome by picosecond time-resolved measurement (Le., pumpprobe) techniques, where we can resolve the induced absorption and its recovery. The picosecond, time-resolved, pumpprobe measurementswere performed over a range of relatively low pump fluences (1 1-64 mJ/cm2) in a 2-mm sample with a linear transmission of -35%. A higher sample concentration, -2 X lo-' M, for the pumpprobe experiments relative to the single-pulse measurements was chosen to improve the signal to noise ratio. The differential transmission, i.e., the change in the probe transmission normalized by the linear transmission ( A T / T ) , is shown as a function of probe delay in Figure 4 for three pump fluences. At each fluence, we observe reverse saturable absorption. Initially the absorption of the solution increases as a function of time in a manner consistent with the temporal integration of the pump pulse. Once the pump pulse has passed through the sample, this initial response is followed by a decay to the original transmission within a few hundred picoseconds. This behavior is consistent with induced absorption in the first excited singlet state that increases as the excited state is populated and then diminishes as the population in that state relaxes to the ground state. Quantitatively, we find that the measured dynamics of the tetramer solution on picosecond time scales are consistent with the physical model described below. Comparison of the experimental results with this model has allowed us to extract values for several system parameters. IV. Theory and Modeling Figure 5 shows the molecular structure of the King's complex as determined from X-ray crystallographyby Neuman et al." The

four iron atoms are equivalent and form a tetrahedron. Each cyclopentadienyl ligand bonds through the five *-orbitals of the carbons to an iron atom. Each carbonyl is pbonded (triply bonding) to three iron atoms on a face of the metal tetrahedron core. This cagelike structure contributes greatly to the molecule's robust nature.

Figure 5. Atomic structureof cyclopentadienyliron carbonyl tetramer after ref 8.

The molecular orbital levels of cyclopentadienylironcarbonyl tetramer were originally determined by Trinh-Toan and coworkers.I2 The highest occupied molecular orbital (HOMO) is predominately composed of essentially nonbonding iron atomic orbitals. The orbital is fully occupied, giving rise to the diamagnetic character of the compound. The lowest unoccupied molecular orbital (LUMO) is formed from six strongly antibonding tetrairon orbitals. The lowest energy transition, the HOMO to the LUMO, is from a nonbonding predominately d-orbital of the irons to an antibonding orbital. This d-d transition should primarily involve the metal core of the cluster and cause a general bond weakening within the tetragonal structure. The ground state is a singlet, and an allowed transition will conserve spin to an excited singlet state. Intersystem crossing will allow a spin change of the excited unpaired electron to give a triplet state. The triplet state will possess a longer lifetime than the excited singlet, since relaxation back to the ground state requires another spin flip transition. Large absorption cross-sections are due to spin-allowed transitions (spin conserved); hence, the excited-state transitions will be either singletsinglet transitions or triplet-triplet transitions. The nature of these excited-state transitions is difficult to predict or to experimentally determine. The molecular orbital diagram suggests three major possibilities. First, the transition could involve promotion of another electron out of the nonbonding metal orbitals to the strongly antibonding metal orbitals. Although this possibility is difficult to exclude, one would expect significant cluster fragmentation from this process, which was not observed. A second possibility for the transition is an excitation from a low-lying fully occupied cyclopentadienyl *-bonding orbital into the now partially occupied metal nonbonding orbital. This transition involves movement of charge from the cyclopentadienyl ligand to an iron atom. This type of transition, referred to as a ligand to metal charge transfer, possesses a large absorption cross-section due to the induced dipole occurring from the charge separation. The final possibility is that the transition is from the newly occupied metal antibonding orbital to a low-lying carbonyl ?r-orbital. This involves movement of charge from the metal core to a carbonyl, resulting in a metal to ligand charge-transfer transition. This transition would also possess a strong absorption cross-section. It is likely that either chargetransfer transition, discussed above, will have a larger absorption cross-section than the d-d groundstate transition. Thus, the presence of such transitions is consistent with an excited state that is more strongly absorbing than the ground state and the observation of RSA. It is not necessary, however, to determine the exact assignment of the excited-state transition for the purpose of modeling our results. It is sufficient only to include these levels in the model. We will attempt to experimentally determine the excited-state transition in future work. The above discussion leads to a simplified five-level model (shown in Figure 6) consisting of three singlet and two triplet states. In this model, level 0 corresponds to the HOMO,while level 1 corresponds to the LUMO, and the ground-state absorption

Allan et al.

6316 The Journal of Physical Chemistry, Vol. 96, No. 15, 1992

2wlw 4 =I 2

TABLE I: Measured Parameters for King's Comolex capture cross-section

branching ratio T

S i e tStab

Triplet States

Figure 6. Proposed five-level model for reverse saturable absorption in King's complex. The symbols are defined in the text of section IV.

is associated with 0-to-1 transitions. Intersystem crossing is associated with the 1-to-3 transition, and excited-state absorption may occur in either the singlet system (1-to-2 transition) or the triplet system (340-4 transition). By inspection of Figure 6, we obtain the coupled differential equations describing the photodynamics of this system:

dN2/dt = UI~INI - N2/721

(2)

dN3/dt = -u34Ih'3 - N3/730 + N4/743

+ N I / T I ~ (3)

- N4/743

(4)

and dNd/dt =

034IN3

Here, I is the incident photon flux (cm-2 d),Ni is the density of molecules in level i, and u and 7 are the respective cross-sections and decay times. The branching ratio Br = T ~ ~ is/ theT ratio ~ ~ of molecules that relax to the ground state versus intersystem crossing to the triplet state. In addition conservation of molecules gives

N = No

+ Nl + N2 + N3 + N4

(5)

The propagation of the pulse is determined from the wave equation, which assuming a slowly varying envelope approximation can be written as dI/dz = -~olINo- u I ~ I N- ~~34IN3

(6)

From the above equations we generate numerical fits to the experimental data and consequently extract the relevant crosssections and decay times. The procedure we follow is to first fit the low fluence data where the population in level 2 may be neglected. We can also neglect N3and N4 as the experimentally measured induced absorption recovers on time scales too short for triplet-state transitions. For these conditions, the system is completely specified by the ground-state cross-section, uOlrthe singlet excited-state cross-section, uI2,and the lifetime of the first excited singlet state, rl0. Moreover, uI2and T~~ are largely decoupled parameters for pulse durations shorter than T ~the ~ peak ; AT/T near zero delay is primarily sensitive to uI2,while the decay of AT/ T determines rl0. The linear absorption measurements provide uol, and fitting to the data allows us to uniquely extract T~~ and uI2. The theoretical curves superposed on the data in Figure 4 were obtained using a single set of parameters, u12= 8.0 X cm2 and T~~ = 120 ps and are in good agreement with the data over the fluence range 11-64 mJ/cm2. At the higher fluences, the induced absorption becomes more sensitive to the upper excited singlet decay time, 721, due to the greater number of excited-state transitions. The numerical fits to the data indicate that the upper excited-state recombination time is considerably shorter than the pulse duration, and we estimate 721I 1 ps. The excellent agreement between the numerical simulations and the data reported here indicates that RSA of picosecond pulses in cyclopentadienyliron carbonyl tetramer is predominantly due to optical transitions within the singlet system. To determine whether the pump-probe results were sensitive in any way to induced anisotropy and reorientational diffusion

recombination time

uol = 4.1 X lo-'* cm2 ui2 = 8 X lo-'* cm2 ~

~

710

= 120 PS

721

< 1 PS

>/ 23T

~

~

in the molecular system, we repeated our measurements with the probe polarized at the "magic angle" l 3 of 54.7O to the pump, at which the probe transmission should be insensitive to reorientational dynamics. We found that both the peak value and the measured decay of AT/ T in this configuration were indistinguishable from that measured with the probe and pump orthogonally polarized. This is consistent with the anticipated highly symmetrical nature of the excited molecules. Previous nanosecond measurementss of the excited-state absorption of the King's complex suggest that for nanosecond optical excitations a long-lived (7% = 60 ns) induced absorption is present. This residual absorption was tentatively attributed to a buildup of population in the triplet state. The overall contribution of the triplet state to the absorption of the pump and probe in the picosecond experiments reported here is difficult to quantify, but it has to be extremely weak. Clearly, if triplet absorption is present, its maximum value is less than our minimum detectable ATIT. This allows us to place an upper limit for the product of the triplet absorption cross-section with the intersystem crossing rate. For times t in the domain 730 > t > tlo > T~~~~~~we can rewrite the branching ratio as u34/713

Aa/71fll(max)

(7)

where Pa = N3u34is the estimated induced absorption coefficient at time t and Nl(max)is the excited singlet population immediately after illumination. We estimate the excited singlet population N,(max) H 1.5 X 10'' cm-3 for a fluence of 64 mJ/cm2 from the maxi" measured induced absorption near zero delay. At delays -600 ps, the measured AT/T (-AaL) is less than 0.01, and for our L = 2 mm cell, the induced absorption, Pa, is less than 0.05 cm-I. For these values, eq 7 gives u 3 4 / ~ l ?< 2.8 X lo-' cm2/s. If we assume that the cross-section for triplet and singlet excited-state absorption are approximately equal, i.e., u34 H uI2,this gives an intersystem crossing time 7 1 3 > 2.8 ns and a branching ratio >23.

V. Conclusions We have demonstrated that the cyclopentadienyliron carbonyl tetramer in solution with methylene chloride exhibits reverse saturable absorption on picosecond time scales. The dynamics of the optically-excited molecules have been measured using picosecond time-resolved absorption at 532 nm and are consistent with a five-level model with the parameters indicated in Table I. We emphasize that the measured singlet excited-state absorption cross-section is only a factor of 2 larger than that of the ground state. This value is too small to explain the nonlinear transmission behavior previously reported8 for nanosecond excitation of nominally identical solutions. Furthermore, our estimate of the product of the intersystem crossing rate and the triplet absorption cross-section indicates that it is unlikely that triplet absorption could fully account for the nanosecond-excitation response. The implication of these results is that the nanosecond measurments are influenced by additional nonlinear mechanisms not accounted for in the present model. A number of areas of investigation are necessary for a complete description of the picosecond photodynamics of reverse saturable absorption of this tetramer. The assignment of the specific transition associated with the singlet excited-state absorption remains to be determined. To this end, we are currently synthesizing various King's complex derivatives. These derivatives have different ligands substituted and will allow us to determine whether the excited-state absorption involves charge transfer from the iron to the carbonyl ligands or from the cyclopentadienyl ligands to the iron atoms. Furthermore, we have not yet unambiguously measured the intersystem crossing rate and the triplet

J. Phys. Chem. 1992,96,6317-6321 absorption cross-section. We anticipate that the triplet population can be enhanced by optical excitation of an equivalent fluence but longer duration compared to the measurements reported here. Under such conditions there is more opportunity for the molecules to undergo intersystem crossing, since molecules that relax back to the ground state can be reexcited during the excitation pulse. This results in an increased population in the triplet. With the singlet parameters known from the present picosecond measurements, time-resolved studies of the triplet absorption for different optical excitation pulse widths should allow direct measurements of the triplet system parameters u34,N3, and ~ 1 3 Such measurements should help to clarify the current discrepancy between the nanosecond and picosecond responses of the King's complex.

6317

References and Notes ( 1 ) Giuliano, C. R.; Hess, L. D. IEEE J . Quantum Electron. 1967,3 ( 8 , Aug), 358-367. (2) Harter, D. J.; Shand, M. L.; Band, Y . B. J . Appl. Phys. 1984,56 (3,

.

Acknowledgment. The work performed a t Hughes Research Laboratories was supported by the Naval Air Development Center under Contract 62269-90-(2-0239. The research at The University of Iowa was supported by the Defense Advanced Research Projects Agency and the Center for Night Vision and Electro-Optics. Registry No. [(CSHs)Fe(CO)]4,12203-87-1.

Aug 1). (3) Hoffman, R. C.; Stetyick, K. A.; Potember, R. S.; McLean, D. G. J . Opt. Soc. Am. B 1989, 6 (4, Apr). (4) Guha, S.; Frazier, C. C.; Chen, W. P.; Porter, P.; Kang, K.; Finberg, S. E. Proc. SPIE-Int. Soc. Opt. Eng. 1989, 1105, 14-19. (5) Coulter, D. R.; Miskowski, V. M.; Perry, J. W.; Wei, T.-H.; Van Stryland, E. W.; Hagan, D. J. Proc. SPIE-Inr. SOC.Opt. Eng. 1989, 1105, 42-5 1. ( 6 ) Harter, D. J.; Band, Y. B.; Ippen, E. P. IEEE J . Quantum Electron.

1985, 21 (8, Aug). (7) King, R. B. Inorg. Chem. 1966,5 (12, Dec), 2227-2230. (8) Tutt, L.; McCahon, S. W.; Klein, M. B. Proceedings of NL090, Hawaii, 1990. (9) White, A. J.; Cunningham, A. J. J . Chem. Educ. 1980, 57 (4). (10) Bock, C. R.; Wrighton, M. S. Inorg. Chem. 1977, 16, 1309. ( 1 1) Neuman, M. A.; Trinh-Toan; Dahl, L. F. J . Am. Chem. SOC.1972, 94 (10, May 17), 3383-3388. (12) Trinh-Toan; Fehlhammer, W. P.; Dhal, L. F. J . Am. Chem. SOC. 1972, 94 (10, May), 3389-3396. (13) von Jena, A,; Lessing, H. E. Berichte der Bumen-Gesellschaft fuer Physikalische Chemie; Verlag Chemie: D-6940 Weinheim, 1979; Vol. 83, pp 181-191.

Effect on Singlet Oxygen Yield of Charge-Transfer Interaction between Fluoranil and Solvent Alexander P. Darmanyan and Christopher S. Foote* Department of Chemistry and Biochemistry, University of California, Las Angeles, California 90024- 1569 (Received: February 12, 1992; In Final Form: April 10, 1992)

The quenching of triplet fluoranil (FA) by molecular oxygen was studied by IR laser fluorimetry and nanosecond laser photolysis in various solvents. In acetonitrile, CCl,, chloroform, and cyclohexane,the charge-transfer (CT) interaction between triplet fluoranil and solvent (Sol.) is insignificant and the quantum yield of triplet formation and the probability of O2(IA ) generation from triplet fluoranil are close to unity. In aromatic solvents, a substantial decrease in the quantum yield (aA-?of 02(1$) formation (limiting value at [O,] a)was observed; for example, in p-xylene the estimated value is 110-2. In the weakly electron-donating solvents, benzene and chlorobenzene, triplet exciplexes with partial charge transfer [3( FAb-.Sol.6+)] and lifetime of 150 ns were observed. The absorption spectra of the triplet exciplexes are slightly broadened compared with the triplet-triplet spectrum of free fluoranil and have broad structureless CT bands in the near IR region. The values of aAm are -0.4-0.5, and the rate constants for quenching by oxygen [(6-8) X IO* M-'s-l ] of the triplet exciplexes are lower than for triplet fluoranil. Molecular oxygen, a strong electron acceptor, probably interacts mainly with the FAb fragment of the triplet exciplex where excitation is localized, leading to efficient 0 2 ( ' $ , production. Fluoranil forms weak CT complexes with aromatic solvents in the ground state, and their direct excitation produces contact ion-radical pairs (CIRP). In benzene and chlorobenzene,the main deactivation channel for the CIRP is intersystem crossing, leading to triplet exciplex formation.

-

-

Introduction There is much current interest in the effect of charge-transfer (CT) interactions on the efficiency of 0 2 ( l A g ) generation (SA) by different sensitizers. Several authors have shown that for sensitizers with high-energy triplet states which are goad electron donors, the energy of the complete CT state (electron transfer from sensitizer to oxygen) lies below that of the triplet and that these sensitizers have low SA C T quenching of triplet sensitizer was suggested to compete with energy transfer. In polar solvents, the energy of the CT state decreases relative to nonpolar solvents and a further decrease in SAo c ~ u r s . ~ * ~ Direct excitation of contact CT complexes between solvent and oxygen leads to 02('$) formation.68 These authors established that the 02(lAg) quantum yield does not depend on the O2concentration and therefore suggested that 02(IAg) is formed on dissociation of the excited CT complexes. However, Ogilby et al. reported the very interesting result that direct excitation of the contact CT complex 3( 1-methylnaphthalene--02),which has high-energy excited states, gives a I-methylnaphthalene triplet with unit probability in cyclohexane; subsequent triplet quenching 0022-365419212096-6317$03.00/0

by molecular oxygen leads to 02('Ag) generation5 The generation of Oz(lAg) by different triplet CT complexes such as triplet exciplexes and triplet ion-radical pairs (3CIRP) has hardly been studied. A recent paper established that triplet exciplexes between excited chloranil and aromatic solvents with a large contribution of locally-excited triplet generate singlet oxygen? The probability of 02(1$)generation by 3CIRPsformed by the quenching of excited chloranil by simple benzene derivatives (anisole, durene, hexamethylbenzene, and 1,2,3-trimethoxybenzene) is lower than that from free chloranil triplet and is in the range 0.05-0.2. We suggested that low SAvalues for 3CIRF% are mainly caused by interaction of oxygen with the anion radical of the pair, causing energy transfer to be relatively inefficient. The present work deals with the effect of CT interaction between fluoranil sensitizer and various solvents on the efficiency of O,(l$) generation.

Experimental Section The absorption spectra and decay kinetics of triplet fluoranil were investigated using nanosecond laser transient absorption 0 1992 American Chemical Society